California Cliffs Warned Us Days Before Collapsing

This morning with the sudden collapse of a sea cliff in southern California that has now claimed at least three lives. Tons of debris fell on beachgoers yesterday in Encinitas, north of San Diego. The calm before the collapse. In July 2026, scientists at the Scripps Institution of Oceanography revealed the results of a four-year study that uncovered something remarkable along California's coastline. Towering coastal cliffs can appear completely still even while subtle movements are building inside the rock. Using highly sensitive underground sensors, researchers detected accelerating ground movement hours and sometimes days before a cliff collapsed. During one major event in Del Mar on April 21st, 2024, the instruments detected warning signs days before roughly 200 tons of material crashed onto the beach. A cliff that looks stable may already be approaching failure and scientists are now learning how to detect its final warning. The California cliff crisis. California's famous coastline is far less stable than it appears. Roughly 70 percent of the California coast consists of eroding coastal cliffs. These bluffs don't just overlook beaches. Highways, railways, homes, businesses, coastal access points, energy facilities, and other critical infrastructure sit on or near them. The danger below can be even more immediate. In August 2019, a section of cliff roughly 30 feet by 25 feet collapsed onto Grandview Beach in Encinitas, killing three women. Other deadly cliff failures have occurred elsewhere in California, including Big Sur, Santa Barbara, Point Reyes, and San Francisco. That tragedy in Encinitas helped drive funding for better coastal monitoring and research into whether scientists could detect a bluff failure before it happened. Instead of only mapping damage after a collapse, researchers wanted to catch the process while it was still unfolding. That became the California Coastal Landslide Early Warning Research Project. The study focused on several locations in northern San Diego County, including Del Mar, Beacons Beach, and San Elijo State Beach, and researchers weren't simply watching these cliffs with cameras. They placed extremely sensitive instruments directly into the ground because the warning they were searching for could be almost impossibly small. The hidden movement. One of the main instruments was a tilt meter. Tilt meters are commonly used to measure extremely small movements associated with earthquake faults. For this project, researchers used them to detect subtle changes in the angle of coastal cliffs. Their accuracy is extraordinary. The instruments can measure ground tilting of about 10 micrometers across a one meter baseline, roughly one-eighth the width of a human hair. Researchers also installed fiber optic extensometers perpendicular to the cliff edge at Del Mar and Beacons Beach. These instruments measure whether the ground is stretching or compressing. They can detect deformation down to approximately one nanometer, or one billionth of a meter. Think about what that means. A beachgoer could look directly at a cliff and see absolutely nothing unusual. No falling rocks, no obvious movement, no dramatic opening in the ground. But underground, the cliff could already be deforming. The team also deployed seismometers, GPS monuments, wave pressure sensors, lidar mapping, and other monitoring technology. Together, these tools gave researchers something they had rarely had before. A real-time view of what a cliff does before it fails. And one crack in Del Mar was about to provide their clearest example. The Del Mar crack. During a maintenance visit at Del Mar, researchers noticed something small. A new crack had appeared near the top of the cliff. It was only about 0.1 inches wide, but it stretched approximately 30 feet parallel to the cliff edge. That was enough to attract attention. Instead of simply photographing the crack and returning later, researchers were able to watch its movement through their instruments. Over the following weeks, the crack widened at approximately 0.015 inches per day. That rate would have been practically invisible to someone standing nearby. But the sensors recorded it. Then, several rain events affected the area. Rainfall matters because water entering a bluff can contribute to instability. At the same time, Pacific waves continuously attacked the cliff from below, eroding its base. So researchers kept watching the data. For weeks, the crack continued moving. Then, on April 19, 2024, something changed. The tilt measurements began accelerating. The cliff wasn't simply moving anymore. According to the researchers' interpretation of the measurements, it was approaching imminent failure. And for the first time in this event, the data became actionable. Coastal managers were notified, but the cliff was still standing. So how much time was left? The 200-ton collapse. The answer came two days later. At approximately 5 a.m. on April 21, 2024, the Del Mar Cliff failed. An estimated 200 tons of material broke away and crashed onto the beach below. But the most important part of this event happened before the collapse. The sensors had detected accelerating deformation. Researchers recognized the signal. And coastal managers were notified on April 19, two days before the failure. That changes how we should think about what appears to be a sudden cliff collapse. The final fall may happen in seconds. But the physical failure process can begin much earlier. The rock creeps. The ground tilts. Cracks widen. Movement accelerates. And eventually, gravity wins. Here on DiscoverTube, we often examine disasters by looking at what scientists discover afterward. This case is different. Scientists were watching the warning before the disaster happened. But one event could still be coincidence. To know whether Del Mar represented something bigger, researchers needed more failures. And they got them. The five warning signs. During the study period, researchers tracked five separate cliff collapses. Two at Del Mar and three at San Elijo State Beach. The failures ranged from relatively small events to collapses involving several hundred cubic yards of material. But they shared one crucial feature. In all five cases, sensors detected cliff movement several hours to days before collapse. That is one of the strongest findings from the four-year investigation. The exact warning wasn't identical at every site. And researchers cannot yet predict every California cliff collapse. But the pattern suggests that some cliffs don't simply fail instantaneously from a perfectly stable state. They deform first. That creates the possibility of identifying a tipping point where ordinary background movement begins transitioning toward failure. The researchers deliberately selected areas where they expected cliff activity could occur. But there was no guarantee that a collapse would actually happen during the monitoring period. Five events did. And those five failures provided the data needed to show that this kind of monitoring could potentially support prediction. So what pushes an already unstable cliff toward that tipping point? One major factor comes from above. The rainfall trigger. Rainfall and waves are two of the primary drivers of California coastal cliff erosion. Waves strike the base of a bluff, removing material and gradually making the cliff steeper. Rainfall and groundwater attack from another direction. Water can enter fractures and permeable layers inside the cliff, contributing to conditions that trigger landslides in already unstable material. To understand that relationship, researchers didn't rely only on underground sensors. From 2022 through 2025, the Scripps team conducted weekly LIDAR mapping along the coastline from Torrey Pines State Beach to Encinitas. Truck-mounted LIDAR and drone photogrammetry created detailed maps showing how cliffs changed over time. Nearby rain gauges recorded daily rainfall. By comparing cliff changes with rainfall conditions, researchers identified preliminary environmental thresholds associated with increased coastal landslide activity. That opens another possibility. Future warning systems might combine direct ground deformation with rainfall thresholds. One system tells researchers what the cliff itself is doing. The other helps reveal the environmental conditions, pushing it But there is a major problem. Not every California cliff behaves exactly the same way, and one beach showed just how difficult building a larger warning network could become. Before we continue, subscribe to Discover Tube if you want more discoveries about the geological forces changing America in real time. Because detecting the warning is one thing. Turning it into a system capable of protecting an entire coastline is much harder. The Beacons Beach Threat. At Beacons Beach in Encinitas, researchers faced a different challenge. This area sits on an intermittently moving complex landslide, and landslides closed the beach trail in May 2022. It looked like exactly the kind of location where high-sensitivity monitoring could be useful. But installing sophisticated equipment along an active California coastline isn't simple. Permitting complications delayed installation of the most sensitive instruments. As a result, the necessary time series data from those sensors wasn't available until 2025. Researchers concluded that Beacons Beach needs a longer observation period before they can determine whether the more expensive, highly sensitive instruments are useful for predicting failures at that specific location. That limitation matters. California doesn't have one uniform cliff. Each location has different geology, groundwater conditions, wave exposure, erosion history, development, and fracture patterns. The study area itself was limited mainly to north San Diego County. Researchers specifically identified other California locations that could benefit from further investigation, including Orange County, Palos Verdes, Santa Barbara County, Big Sur, Santa Cruz, and Pacifica. So even after four years, scientists are only beginning to understand how widely these warning signals can be applied. And strangely, part of the technology helping them do it comes from another geological hazard entirely. The earthquake technology. Some of the instruments beneath these cliffs were developed from technology used to study earthquakes. Scripps geophysicist Mark Zumberg has spent decades developing instruments capable of measuring extremely precise earth movements. For years, scientists have attempted to understand whether subtle changes around faults could reveal something about future earthquakes. Predicting earthquakes remains extraordinarily difficult, but the California Cliff Project offered researchers another opportunity. Could instruments designed to detect microscopic earth movement identify the deformation that occurs before a landslide? The answer from these monitored events was promising. Tilt meters detected tiny angular changes. Extensometers recorded stretching and compression. Seismometers monitored shaking within cliff faces. GPS monuments measured changes in cliff top position. Wave pressure sensors measured the force reaching the cliff base. And LIDAR repeatedly mapped the physical coastline. Instead of relying on one signal, researchers could begin comparing multiple measurements. That matters because a useful warning system can't simply detect movement. It has to distinguish dangerous movement from normal background change. How much acceleration is too much? How much rainfall changes the risk? How long after a threshold is crossed could a cliff remain standing? Those questions are now becoming the next phase of the research. And if scientists can answer them, the result could move beyond geological monitoring. It could become a public warning system. The California Warning System. The July 2026 report concluded that in-ground sensors can provide warnings of dangerous landslides with hours to days of notice at the studied sites. But researchers also emphasized that more work is required before those findings become a formal, actionable system. That next step isn't just scientific, it's operational. What happens when a sensor detects imminent failure? Who receives the warning? Who decides whether a beach closes? How quickly can authorities respond? The report recommends creating protocols involving coastal managers, lifeguards, marine safety units, rail operators, and other stakeholders. Possible responses could include temporary beach closures, closing areas above the cliff, adding warning signs, stabilizing vulnerable sections, or in some situations, carrying out controlled removal of material likely to fail. The Del Mar event shows why those protocols matter. Researchers had data suggesting failure was approaching. They contacted coastal managers. Two days later, the cliff came down. Now imagine that same process operating across multiple high-risk locations. Could California eventually monitor its most dangerous bluffs continuously? And could an automated signal one day trigger action before people even realize anything is moving? That's the larger possibility behind this research. But there's another piece of the problem that is easy to overlook. Even if scientists know a cliff is dangerous, they still need to know how exposed the people below it are. The beach danger zone. A collapsing cliff isn't equally dangerous under every condition. Beach width matters. At low tide, people may have a broad area of sand separating them from the bluff. At another time, waves and tides may force beachgoers much closer to the cliff. As part of the research, the team developed an experimental smartphone app at Torrey Pines State Beach. It wasn't an app predicting cliff collapses. Instead, it forecast beach width. Using the latest LiDAR beach survey, along with wave and tide forecasts, the experimental system produced hourly beach width predictions up to three days in advance. Why does that matter? Because knowing when a beach becomes narrow can help determine when people have less room to stay safely away from unstable cliffs. On DiscoverTube, this is where the story becomes larger than one sensor or one landslide. The goal isn't simply predicting geology. It's translating geological data into information that can actually reduce risk. Ground sensors could reveal instability. Rainfall thresholds could indicate elevated landslide conditions. LiDAR could show how the cliff is physically changing. And beach forecasts could reveal how exposed visitors may be. Combine those layers and scientists begin moving from observation toward prevention. But several critical questions remain unanswered. The next collapse. Researchers have demonstrated that prediction is possible under the conditions they studied. They have not demonstrated that every cliff collapse along California can now be predicted. That distinction is important. Scientists still need to determine the best combination of sensors, coverage, sensitivity, and cost. Some of the largest landslides occur infrequently. That means researchers may need years of additional observations before enough events occur to understand the full range of warning behavior. And then there is the false alarm problem. Imagine closing a beach because sensors detect unusual movement and nothing happens. Do it once and people understand. Do it repeatedly and warnings could lose credibility. But the opposite mistake could be far worse. What if scientists underestimate a signal and the cliff fails while people are standing below? Building an effective warning system means finding the line between those two outcomes. That's why more data matters. More collapses. More rainfall events. More locations. More years of measurements. The existing research provides evidence that the warning exists. The next challenge is learning how reliably scientists can recognize it. And that brings us back to the most dramatic event in the entire study. April 19, 2024. The cliff was still standing. Beachgoers looking at it might have seen nothing extraordinary. But beneath the surface, instruments detected acceleration. Researchers believed failure was approaching. They warned coastal managers. Then, on April 21, approximately 200 tons came down. If scientists hadn't been measuring the cliff, that warning would have remained invisible. How many other California cliffs could be sending similar signals right now? The final warning. The calm before the collapse may not actually be calm. On July 9, 2026, Scripps Institution of Oceanography released the results of a four-year investigation showing that some coastal cliff failures can produce measurable warning signs hours to days before collapse. Five collapses were tracked. Two occurred at Del Mar. Three occurred at San Elijo State Beach. And in every one of those monitored failures, sensors detected movement beforehand. At Del Mar, researchers watched a small crack widen at approximately 0.015 inches per day. Then the movement accelerated. Coastal managers were notified on April 19. Two days later, at approximately 5 o'clock in the morning, an estimated 200 tons of cliff material crashed onto the beach. For DiscoverTube, this is what makes the discovery so important. Scientists aren't claiming California's entire coastline can suddenly be predicted. They're showing that a collapse that looks instantaneous to us may actually have a measurable beginning. A movement smaller than a human hair. A crack slowly widening. A tiny change in tilt. Then acceleration. Then failure. Researchers now want longer observations and monitoring at additional California locations. And if the same warning patterns continue appearing, California could eventually have something coastal communities have never had before. Time. Not weeks. Not necessarily days. Sometimes perhaps only hours. But when people are standing beneath hundreds of tons of unstable rock, even hours could matter. So here's the question. Should California expand this sensor network along its most dangerous coastal cliffs? Or should the technology first be proven across more locations? Tell us what you think in the comments below. And if you want more stories about the hidden geological changes scientists are detecting across America, subscribe to DiscoverTube. Because sometimes the biggest disaster isn't the one that comes without warning. It's the warning we didn't know how to see.